Bugfix update VNL-ATK 2016.3

Download it from http://quantumwise.com/products/download as usual.

For more information about VNL-ATK 2016, see the original release letter.

Please note that for this release, we are introducing a new end-user license agreement (EULA).

2016.3 updates

Important bugs or improvements, possibly affecting results or output, or hindering usage of the software

• IETS analyzer: the total d2I/dV2 signal should be divided by the total dI/dV signal (not by the inelastic dI/dV signal, as it was)
• Fix for VNL crash when showing density plots of device configuration
• Avoid deadlocks in checkpoint files when running in parallel
• Added terms to be able to correctly calculate forces with the Ozaki contour integration method
• Save the transmission spectrum while running IVCurve
• Fix for the doping plugin, to ensure it always writes tags and doping to electrodes

Regular bugs

• HamiltonianDerivatives, fixed check for atomsTranslatedOutsideUnitCellInDirection for hexagonal cells
• Fixed ammonia in the fragments in the MolecularBuilder (had four H atoms)
• Fixed the determination of the Brillouin zone (Wigner-Seitz cell) for high-aspect lattices
• DynamicalMatrix: clean up large amount of empty log files when using a classical potential
• Avoid degenerate unit cells in Interface Builder
• Fix for DiagonalizationSolver, could hit communicator limit in MPI
• AKMC: fix numbering mismatch between the log file and the filenames.
• Ensure animated GIF can be exported also when "wrap atoms" is used
• Molecular builder would throw an exception if stash was empty
• Enable create button when in preview mode (SubstitutionalAlloyPlugin)
• Fix for cut planes which lose resolution when repeated in viewer
• Keyword "enable_symmetries" only works in ATK 2016 is the crystal is oriented in the standard way (for instance, hexagonal crystals must have the C axis along Z); will be fixed more generally for 2017, but at least a warning has been added in 2016.3
• Avoid adding redundant script blocks when they are read from analyis from file (in particular for dynamical matrix)
• Make sure tag selection tool for trajectories in Viewer doesn't raise an exception
• Fix exception thrown when moving splitter in MovieTool
• Fix for using Polycrystal Builder with MoleculeConfiguration
• Fix for problem with zero-dimensional data in ElectronPhononCoupling analyzer

Specific enchancements and bug fixes for SurfaceConfiguration

• Improved "surface from bulk" to recognize shorter repetition patterns
• Corrected how doping is handled for surfaces configurations
• Fix a segfault in the translate plugin when used for SurfaceConfiguration
• Fix translate plugin for surfaces
• Overlapping atoms were not detected for SurfaceConfiguration
• Fix passivation of surfaces (also for devices)

Performance improvements

• Don't use fully dense onsite spin-orbit matrices in tight-binding
• For unpolarized calculations, do not evaluate() the DOS twice unnecessarily
• AKMC: Make write_searches default to False when used with a TremoloXCalculator (logging was too slow)

ElectronPhononCoupling

• Significant improvements of performance (memory and speed, sometimes from 200 to 2 GB)
  • Restructure tasks to avoid large (k,q)-lists
  • Always calculate E(k) and v(k) for all initial states
  • Do not run empty task lists
  • Enabling sparse storage of energies and velocities
• Changed default value of energy_tolerance
• Improved legacy support for sparse/dense energies
• Enable sparse energies to be shown in ElectronPhononCoupling analyzer

Mobility

• Make calculation of Hall coefficients optional - takes a long time with many k-points
• Legacy support + unittest. Early returns if the coupling matrix is zero
• Don't include states with zero inverse relaxation times
• Allow for selecting only some electron bands

General enhancements

• GUI support for RegularKPoint and AdaptiveKPoint grid for IVCurve
• Add nlprint support to MDTrajectory
• Do not allow installing VNL-ATK in a path containing non-ascii characters (Python 2.x does not support this)
• Support reading OUTCAR files generated with IBRION=7

Smaller improvements, bugs, or technical issues, incl. docstring and manual

• Disable repetition of surfaces in Z-direction in Viewer
• Properly handle NewCalculator if no potentials for ATKClassical
• Cleaver, improved tab sequence / initial focus
• Set an upper bound on how many k-points can be specified to avoid Script Generator becoming slow or segfaulting
• Make sure electrodes are moved when reading a Trajectory for devices if the central region expands/shrinks
• Preview overlay text is now back again in the Builder
• 'Any files (*)' as the only option for the chemistry importer
• The Job Manager should not accept dropped nc files.
• Added a progress export when exporting a GIF for an MDTrajectory
• Set lower bound on temperature in Markov Chain Analyzer
• Fix indentation in optimize geometry script
• Fix zoom issue in Optical Spectrum Analyzer
• Problem in MD when using MoleculeConfiguration with hook funtions fixed
• Fixed typo in Nanowire plugin
• Disable "Adjust configuration" when analysis from file is used
• CustomScripter, small issue about filename
• Fix typo in QuickOptimizer plugin
• Make last downloaded item active in COD Database
• Set a lower bound on forces in QuickOptimizer plugin
• Fix for comboboxes in VibrationalVisualizer
• Better message when attempting to use a network license file as standalone
• Change version of nc files from 2017 to 2016.
• Weird text behind scrollbar in Builder
• Fix AddOn Manager hiding behind main window
• Tone down the warning in depth fade window if no configuration
• Rephrased the setting "Use local plugins" to make it more understandable
• Fixed consistency of Isoline checkbox in cutplane properties
• Remove html tags from COD Database results
• Clamp axis limits to >1e-16 in log-scale.
• Give a proper error message when HTSTEvent is given SurfaceConfigurations.